[ntp:questions] Re: how many servers should I poll?

mike cook michael.cook at wanadoo.fr
Fri Aug 8 09:49:00 UTC 2003


Think of it as trying to get a good approximation to 'true time' with
ordinary watches. They, like computer oscillators are always going to be 
fast or slow to varying degrees.

Let's say you start off the watch in sync with the news pips.

Now wait a few days.
A.  If you only have one watch, you know what the time is, but you do
     not know if it is correct.
B.  If you have two watches you don't know what the time is period.
C.  If you have 3, You can have a pretty good idea of the correct time
     until one breaks, then you are in situation B.
D.  If you have four watches, you can a sure of having a good idea
     of the time even if one breaks.

As to stratum, I have a couple of DCF radioclocks (situation B) and
back that up with 3 pool.ntp.org servers (situation D+)


  remote       refid          st t when poll reach  delay  offset jitter
  ====================================================================
  127.127.1.0 127.127.1.0      10 l   6   64  377   0.000   0.000   0.004
+127.127.28.0  .DCF.           0 l   5   64  377   0.000  -0.046   0.166
*127.127.28.1  .DCF.           0 l   6   64  377   0.000   1.269   0.050
  192.168.1.255 0.0.0.0        16 -   -   64    0   0.000   0.000 4000.00
-64.44.160.38  128.252.19.1    2 u   3   64  377 163.757   7.978   0.943
+204.17.42.198 207.126.98.204  2 u  61   64  377 219.236   4.844   1.455
-205.142.198.4 130.207.244.240 2 u   7   64  377 163.190   8.328   1.373








More information about the questions mailing list